Cost of Quality and Theory of Constraints: A Management Guide

Quality costs fall into four categories, and the first discipline is sorting every quality-related dirham into the right one. Prevention costs stop defects happening: quality engineering, operator training, preventive maintenance, supplier quality development, mistake-proofing (poka-yoke) devices. Appraisal costs find defects: incoming inspection, in-process testing, final inspection, lab equipment, quality audits. Internal failure costs are defects caught before the customer: scrap, rework, re-inspection, downtime, downgraded product. External failure costs are defects that reach the customer: warranty claims, returns, recalls, complaint handling, and the unrecorded killer — lost customers and reputation. The running example is Falcon Aluminium, a JAFZA extruder. Its annual quality ledger: prevention AED 400,000 (training, preventive maintenance, supplier audits), appraisal AED 600,000 (testing lab, in-process inspection), internal failure AED 1,800,000 (scrap, rework, downgrades), external failure AED 2,200,000 (warranty, replacements, a major customer claim) — total cost of quality AED 5.0 million on AED 60 million sales, 8.3%. The mix is the diagnosis: failure costs (AED 4.0 million) dwarf prevention and appraisal (AED 1.0 million) — the classic profile of a company under-investing in getting it right first time.

Total quality cost is the sum of control costs (prevention + appraisal, rising with quality effort) and failure costs (falling with quality effort). The economic optimum — the quality level minimising total cost — is where the marginal dirham of prevention saves exactly a dirham of failure. Spend less and failure costs swamp the savings; spend more and prevention becomes gold-plating. For Falcon, the numbers suggest it is far left of the optimum: shifting AED 300,000 from failure to prevention — say, a press-maintenance programme and operator certification — that cuts internal failure by AED 500,000 and external failure by AED 400,000 would save AED 600,000 net. The empirical regularity behind this: external failure costs typically run multiples of internal failure (a defect caught at final inspection costs scrap value; the same defect in a customer's curtain-wall system costs replacement, labour, and the relationship), and internal failure multiples of prevention. So the practical rule for most companies: the optimum is well above current prevention spending — keep investing in prevention until the measured failure reduction per dirham falls to one. Two caveats: some prevention spending has long paybacks (culture change, supplier development) that annual budgets punish — fund them as investments with multi-year tracking; and in regulated or safety-critical work, the optimum computed from recorded costs understates the true external failure (liability, licence risk), so spend beyond the naive optimum.

Quality costs hide across the accounts — scrap in cost of goods sold, warranty in sales expenses, inspection in production labour, training in HR — so the first reporting task is extraction. Falcon's quarterly cost-of-quality report presents the four categories in dirhams and as percentages of sales, trended over eight quarters, with the failure categories broken by cause (press-related, finishing-related, supplier material). The report goes to the operations meeting, not just finance, because every category has an operational owner. Three reporting disciplines make it work: trend, don't snapshot — a single quarter's 8.3% means little; eight quarters falling from 11% to 8.3% with prevention rising means the programme is working; benchmark the mix, not just the total — best-in-class manufacturers run total COQ at 2–4% of sales with prevention exceeding failure, a target that tells Falcon how far it has to go; and report external failure honestly including estimates of lost business — the customer who quietly moved to a competitor after the curtain-wall claim never filed a warranty claim, but the sales ledger knows. The report's political function matters as much as its arithmetic: it converts quality from an engineering slogan into a P&L line that the CFO defends in budget season.

Eliyahu Goldratt's Theory of Constraints starts from an observation every plant manager recognises: improvements everywhere do not improve the whole. Falcon's extrusion line has four stages — billet heating, extrusion press, anodising, packing — with hourly capacities of 10, 6, 8 and 12 tonnes. The press, at 6 tonnes an hour, is the constraint: the system can never exceed 6 tonnes an hour, so an hour lost at the press is 6 tonnes of throughput lost forever, while an hour saved at packing (capacity 12) gains nothing — the packing team just waits longer. The five focusing steps: 1) Identify — the press is the constraint (capacity analysis, or simply the workstation with the permanent queue before it). 2) Exploit — maximise the press's productive time without spending: schedule to minimise die-change downtime, ensure billets always queue before it (never starve the constraint), inspect billets before pressing so press time is never wasted on defective input. 3) Subordinate — everything else dances to the press's rhythm: heating prepares exactly what the press needs, anodising and packing flex around its output; non-constraints carry protective buffer, not maximum utilisation. 4) Elevate — when exploitation is exhausted, invest: a second press, faster die changes, outsourced overflow. 5) Repeat — the constraint will move (to anodising, to the market itself); go back to step 1, and do not let yesterday's solution become today's inertia. The common failure is step 3: managers cannot bear to see expensive non-constraint equipment idle, so they keep it busy building inventory — which is waste wearing the disguise of efficiency.

Throughput accounting redefines the P&L for decision-making: throughput (T) = sales − totally variable costs (in practice, direct materials — the only cost truly variable with each unit in the short run); investment (I) = inventory and equipment; operating expense (OE) = everything else, treated as fixed for the decision horizon; net profit = T − OE. Product-mix decisions then rank products by throughput per constraint minute. Falcon's two products: standard profiles — price AED 8,000/tonne, materials AED 4,500, throughput AED 3,500, press time 10 minutes/tonne (6 tonnes/hour) → AED 350 per press-minute; architectural profiles — price AED 12,000, materials AED 5,500, throughput AED 6,500, press time 25 minutes/tonne → AED 260 per press-minute. Throughput accounting says: fill the press with standard profiles first (AED 350/minute beats AED 260/minute). Traditional absorption costing might say the opposite — architectural profiles carry higher margins after overhead allocation — but the overhead is fixed regardless, and the press minutes are the scarce resource. With 6 tonnes/hour × 16 hours × 26 days of press time a month, every minute allocated to the lower-throughput product destroys AED 90 of throughput. The method's boundary: it is a short-run decision tool — in the long run, operating expenses are not fixed, capacity can change, and the product mix must cover total costs. Use throughput ranking for the monthly mix; use full costing for the strategic question of whether the product portfolio covers everything.

The two halves of this page combine into one operational rule: quality effort should be concentrated at and before the bottleneck. A defective billet pressed for 10 minutes and then scrapped at final inspection wasted 10 minutes of constraint time — time that can never be recovered, worth AED 3,500 of throughput per tonne. The same defect caught by incoming billet inspection cost only appraisal time at a non-constraint. So Falcon's quality plan under TOC: prevention spending targeted at press-related causes (die maintenance, billet quality — the press's inputs), appraisal positioned before the press (billet inspection, first-off checks immediately after die changes), and rework routed to never consume press time (rework on non-constraint equipment, or scrapped rather than re-pressed when press time is scarce). The cost-of-quality report gains a constraint dimension: failure costs split into 'consumed constraint time' versus 'non-constraint', because a dirham of failure at the bottleneck costs multiples of a dirham elsewhere. This is where management accounting earns its keep — not in computing the numbers, but in pointing the improvement effort at the one place where it moves the whole system's profit.

The five focusing steps need a scheduling method, and TOC's is drum-buffer-rope. The drum is the constraint's schedule — the press's production plan sets the beat for the whole plant, because nothing else can go faster. The buffer is protective inventory (or time) placed before the constraint — enough billets queued before the press that a hiccup upstream never starves it, but no more, because excess buffer is working capital earning nothing. The rope is the signal tying material release to the constraint's pace — billets are released into heating at the rate the press consumes them, not at the rate heating could produce. Falcon's implementation: the press schedule (the drum) is built from the throughput-ranked product mix; a two-shift billet buffer sits before the press sized from heating's variability; and the rope is a simple kanban — heating starts a new batch only when the buffer drops below its trigger. Contrast this with conventional MRP, which schedules every workstation to its own capacity and optimises locally: drum-buffer-rope deliberately leaves non-constraints unscheduled in detail, because their only job is to serve the drum. The accounting consequence: work-in-progress falls (the rope stops over-release), the buffer's carrying cost is visible and sized deliberately, and the plant's delivery performance becomes a function of buffer management — missed due dates trace to buffer penetration, and expediting means protecting the buffer, not shouting at every workstation.

TOC measures the system with three numbers: throughput (T) — the rate at which the system generates money through sales; inventory/investment (I) — all money tied up in things the system intends to sell; operating expense (OE) — all money spent turning inventory into throughput. From these: net profit = T − OE; return on investment = (T − OE) / I; productivity = T / OE. Notice what is missing: efficiency metrics per workstation. Traditional measures — labour efficiency, machine utilisation at non-constraints — are not just useless under TOC, they are harmful: a 95% utilisation target at packing guarantees overproduction. Falcon's monthly TOC dashboard shows: throughput AED 4.2 million, OE AED 3.1 million, constraint (press) utilisation 94%, buffer penetration events 3, on-time delivery 97%. The press utilisation is the only utilisation reported — non-constraint utilisation is deliberately unmeasured, because measuring it creates pressure to misuse it. When the constraint moves to the market (capacity exceeds demand — the common mature-business state), throughput thinking still applies: the constraint is customer orders, and the five steps redirect to marketing and sales — exploit (serve the best orders first), subordinate (production to demand), elevate (grow demand). The measures adapt: throughput per sales visit, per campaign dirham. TOC's critics note it underplays long-run cost structure — treating OE as fixed forever is a decision-simplification, not a fact — which is why the page pairs it with full costing for strategic questions. But for running the system month to month, T, I and OE keep every eye on the constraint, which is exactly where eyes belong.

Quality costs are failure costs made visible — the spoilage page handles their accounting mechanics. Throughput thinking extends the relevant-costing page's product-mix logic, and the balanced-scorecard page shows where quality and constraint metrics sit in strategy.